Loss of the tumor suppressor phosphatase and tensin homolog (PTEN), the key negative regulator of Phosphatidylinositol 3-Kinase (PI3K) activity, is one of the most common genetic events in primary prostate cancer (PCa). Moreover, its frequency further increases in metastatic PCa, indicating that hyperactivation of the PI3K pathway plays an important role in the pathogenesis of PCa and implicating PI3K as an attractive target in this tumor type. However, early clinical trials with PI3K inhibitors in multipe cancer types, including PCa, have been disappointing. The clinical efficacy of these early PI3K inhibitors, which were largely pan-PI3K inhibitors that block the action of all Class I PI3K isoforms, may have been limited by their lack of isoform specificity. Indeed, preclinical work and emerging clinical trials suggest that inhibitors of individual isoforms may achieve greater efficac with fewer side effects. For example, Cal101, an inhibitor of the p110d isoform of PI3K, has demonstrated remarkable clinical efficacy in certain B-cell malignancies. Early unpublished clinical results also suggest that p110a inhibitors are outperforming pan-PI3K inhibitors in luminal breast cancer. Because the PIK3CA gene encoding p110a is frequently mutated in tumors, p110a has garnered the bulk of the attention from pharmaceutical and basic researchers. However, using a murine genetic model we identified p110 as a key target in PTEN-null prostate tumors. We and others have subsequently shown that human PTEN-null prostate cancer cell lines are selectively dependent on p110. We have further identified Kin-193, also known as AZD6482, as a potent and specific p110 inhibitor suitable for studies in vitro and in mice. Notably, a new p110 inhibitor, GSK2636771, is now in clinical trials in patients with PTEN-deficient advanced solid tumors (NCT01458067). Therefore, it is both imperative and timely to carefully investigate p110 inhibitors in preclinical settings. In this gant we propose to evaluate the therapeutic potential of this novel class of inhibitors in prostate cancer in vitro and in vivo using human cancer cell lines, genetic mouse models as well as primary human prostate tumor explants, to develop biomarkers predictive of response to p110 inhibition, and to identify optimal combination partner agents for p110-based therapy. The studies proposed in this grant will likely provide important information that will help to optimize the clinical impact of this class of inhibitors in PTEN-deficient tumors.